SeMPI: a genome-based secondary metabolite prediction and identification web server

Area of Science:

• Microbiology
• Biochemistry
• Bioinformatics

Background:

• Secondary metabolism in bacteria, fungi, and plants produces numerous bioactive compounds.
• Genome mining efficiently identifies gene clusters, but encoding gene clusters for known natural products remain unidentified.
• Structural elucidation of secondary metabolites is challenging due to unpredictable post-modifications.

Purpose of the Study:

• To introduce SeMPI, a web server for predicting and identifying natural products synthesized by type I modular polyketide synthases (PKS).
• To improve the identification of biosynthetic gene clusters and the structural elucidation of secondary metabolites.

Main Methods:

• Utilized genome mining for identifying gene clusters.
• Implemented a structural comparison with annotated natural products to refine PKS product structure predictions and include putative tailoring reactions.
• Developed a benchmark dataset using 40 gene clusters with annotated PKS products.

Main Results:

• SeMPI provides a pipeline for predicting natural products synthesized by PKS.
• The web server integrates structural comparisons to enhance prediction accuracy and account for post-modifications.
• A benchmark demonstrated the pipeline's utility.

Conclusions:

• SeMPI offers an efficient tool for identifying natural products from PKS gene clusters.
• The approach aids in discovering novel bioactive substances by bridging genome data and chemical structures.
• The SeMPI web server is publicly accessible for research.